It's not a good image for GitHub and authors when they make a dedicated post about performance with "fast and performant" in the headline, even boast a 10pp CPU usage difference, when:
1) it gets close to max CPU usage, and 2) still stutters, even on modern processors; with hardware acceleration disabled.
This was raised in previous submissions on HN, and apparently even acknowledged, but I assume they completely dismiss this to the point of not even bothering (as far as I can tell) to test acceleration support in the browser.
I would put all hardware in the same bag - can be good as new after 5+ years - but whatever electronics you want to buy today (be it a phone, "smart" TV, even a car), if you have any experience you know that vendors are now mostly in the business of providing a service, not just a physical product. And that's what you'll want to look up and pay most for. If the stuff becomes incompatible or not secure for everyday use because the vendor stopped (or didn't even intend) to release updates, what's the good?
Thankfully, it's still possible to update software yourself. If that's impossible, well, good bye from me. I'm not going to be forced to buy new stuff because the perfectly good old one is not going to be updated.
Quick glance: there doesn't appear to be a way to link to specific messages for verification - I understand it will be difficult for large files, etc. but you can probably squeeze a base64-encoded value in the URL #hash for short messages.
Also HN, are there any other existing services for proofs?
When you submit a message, I locally generate a file you can download (or copy to clipboard), that contains the message, hash and the release it will go to.
For text messages, I could also generate a special link. Good idea!
Edit: to follow up on your other question, I found a bunch of other services like that. Most of them are paid per proof, though.
This was raised in previous submissions on HN, and apparently even acknowledged, but I assume they completely dismiss this to the point of not even bothering (as far as I can tell) to test acceleration support in the browser.